Active Inference Research Articles

Active Inference and Artificial Spin Ice: Control Processes and State Selection.

Theory and simulations are used to demonstrate implementation of a variational Bayes algorithm called "active inference" in interacting arrays of nanomagnetic elements. The algorithm requires stochastic elements, and a simplified model based on a magnetic artificial spin ice geometry is used to illustrate how nanomagnets can generate the required random dynamics. Examples of tracking and PID control are demonstrated and shown to be consistent with the original stochastic differential equation formulation of active inference. Interestingly, nonlinear response in the form of spikes and spike trains not predicted by the original theory can appear in the nanomagnet system for certain temperature regimes. A theoretical approach using a mean-field approximation for spin systems is proposed, which describes the transition to nonlinear response. Finally, the possibility to create simple magnetic arrays using realistic models is shown with micromagnetic simulations of a simple 17 element array of nanomagnets that include magnetic anisotropies, and exchange and dipolar interactions. Possible applications are simulated to illustrate how nanomagnetic arrays can be used as the stochastic element for feedback control of processes, investigation and control of magnetic state evolution, and as a method to optimize pulsed field magnetic switching protocols.

Introducing ActiveInference.jl: A Julia Library for Simulation and Parameter Estimation with Active Inference Models

We introduce a new software package for the Julia programming language, the library ActiveInference.jl. To make active inference agents with Partially Observable Markov Decision Process (POMDP) generative models available to the growing research community using Julia, we re-implemented the pymdp library for Python. ActiveInference.jl is compatible with cutting-edge Julia libraries designed for cognitive and behavioural modelling, as it is used in computational psychiatry, cognitive science and neuroscience. This means that POMDP active inference models can now be easily fit to empirically observed behaviour using sampling, as well as variational methods. In this article, we show how ActiveInference.jl makes building POMDP active inference models straightforward, and how it enables researchers to use them for simulation, as well as fitting them to data or performing a model comparison.

Flow and intuition: a systems neuroscience comparison.

This paper explores the relationship between intuition and flow from a neurodynamics perspective. Flow and intuition represent two cognitive phenomena rooted in nonconscious information processing; however, there are clear differences in both their phenomenal characteristics and, more broadly, their contribution to action and cognition. We propose, extrapolating from dual processing theory, that intuition serves as a rapid, nonconscious decision-making process, while flow facilitates this process in action, achieving optimal cognitive control and performance without [conscious] deliberation. By exploring these points of convergence between flow and intuition, we also attempt to reconcile the apparent paradox of the presence of enhanced intuition in flow, which is also a state of heightened cognitive control. To do so, we utilize a revised dual-processing framework, which allows us to productively align and differentiate flow and intuition (including intuition in flow). Furthermore, we draw on recent work examining flow from an active inference perspective. Our account not only heightens understanding of human cognition and consciousness, but also raises new questions for future research, aiming to deepen our comprehension of how flow and intuition can be harnessed to elevate human performance and wellbeing.

Human-Robot Kinaesthetic Interactions Based on the Free-Energy Principle

Human-Robot Kinaesthetic Interactions Based on the Free-Energy Principle

Leveraging Explainable AI and Multimodal Data for Stress Level Prediction in Mental Health Diagnostics

The increasing prevalence of mental health issues, particularly stress, has necessitated the development of data-driven, interpretable machine learning models for early detection and intervention. This study leverages multimodal data, including activity levels, perceived stress scores (PSS), and event counts, to predict stress levels among individuals. A series of models, including Logistic Regression, Random Forest, Gradient Boosting, and Neural Networks, were evaluated for their predictive performance. Results demonstrated that ensemble models, particularly Random Forest and Gradient Boosting, performed significantly better compared to Logistic Regression. Random Forest achieved an accuracy of 73%, while Gradient Boosting delivered a balanced precision-recall tradeoff with an accuracy of 72%. Gradient Boosting outperformed in identifying high-stress instances, achieving a recall of 70%, making it the most reliable model for stress prediction. Explainable AI (XAI), which refers to the ability of machine learning models to provide transparent and understandable explanations for their predictions, was employed in this study using SHAP (SHapley Additive exPlanations). SHAP values revealed that Perceived Stress Score (PSS) had the most significant impact on predictions, followed by event count and activity inference. Higher PSS scores strongly correlated with high-stress predictions, while increased event counts and activity levels were associated with lower stress. These findings underscore the importance of incorporating behavioral patterns in stress diagnostics and highlight the utility of explainable models in improving transparency, trust, and usability for clinical decision-making. This research establishes a robust foundation for deploying interpretable machine learning systems to support mental health diagnostics and enhance clinician decision support.

Realizing Synthetic Active Inference Agents, Part II: Variational Message Updates.

The free energy principle (FEP) describes (biological) agents as minimizing a variational free energy (FE) with respect to a generative model of their environment. Active inference (AIF) is a corollary of the FEP that describes how agents explore and exploit their environment by minimizing an expected FE objective. In two related papers, we describe a scalable, epistemic approach to synthetic AIF by message passing on free-form Forney-style factor graphs (FFGs). A companion paper (part I of this article; Koudahl etal., 2023) introduces a constrained FFG (CFFG) notation that visually represents (generalized) FE objectives for AIF. This article (part II) derives message-passing algorithms that minimize (generalized) FE objectives on a CFFG by variational calculus. A comparison between simulated Bethe and generalized FE agents illustrates how the message-passing approach to synthetic AIF induces epistemic behavior on a T-maze navigation task. Extension of the T-maze simulation to learning goal statistics and a multiagent bargaining setting illustrate how this approach encourages reuse of nodes and updates in alternative settings. With a full message-passing account of synthetic AIF agents, it becomes possible to derive and reuse message updates across models and move closer to industrial applications of synthetic AIF.

Learning dynamic cognitive map with autonomous navigation.

Inspired by animal navigation strategies, we introduce a novel computational model to navigate and map a space rooted in biologically inspired principles. Animals exhibit extraordinary navigation prowess, harnessing memory, imagination, and strategic decision-making to traverse complex and aliased environments adeptly. Our model aims to replicate these capabilities by incorporating a dynamically expanding cognitive map over predicted poses within an active inference framework, enhancing our agent's generative model plasticity to novelty and environmental changes. Through structure learning and active inference navigation, our model demonstrates efficient exploration and exploitation, dynamically expanding its model capacity in response to anticipated novel un-visited locations and updating the map given new evidence contradicting previous beliefs. Comparative analyses in mini-grid environments with the clone-structured cognitive graph model (CSCG), which shares similar objectives, highlight our model's ability to rapidly learn environmental structures within a single episode, with minimal navigation overlap. Our model achieves this without prior knowledge of observation and world dimensions, underscoring its robustness and efficacy in navigating intricate environments.

Integrative multi-omics analysis reveals a novel subtype of hepatocellular carcinoma with biological and clinical relevance.

Hepatocellular carcinoma (HCC) is a highly heterogeneous tumor, and the development of accurate predictive models for prognosis and drug sensitivity remains challenging. We integrated laboratory data and public cohorts to conduct a multi-omics analysis of HCC, which included bulk RNA sequencing, proteomic analysis, single-cell RNA sequencing (scRNA-seq), spatial transcriptomics sequencing (ST-seq), and genome sequencing. We constructed a tumor purity (TP) and tumor microenvironment (TME) prognostic risk model. Proteomic analysis validated the TP-TME-related signatures. Joint analysis of scRNA-seq and ST-seq revealed characteristic clusters associated with TP high-risk subtypes, and immunohistochemistry confirmed the expression of key genes. We conducted functional enrichment analysis, transcription factor activity inference, cell-cell interaction, drug efficacy analysis, and mutation information analysis to identify a novel subtype of HCC. Our analyses constructed a robust HCC prognostic risk prediction model. The patients with TP-TME high-risk subtypes predominantly exhibit hypoxia and activation of the Wnt/beta-catenin, Notch, and TGF-beta signaling pathways. Furthermore, we identified a novel subtype, XPO1+Epithelial. This subtype expresses signatures of the TP risk subtype and aligns with the biological behavior of high-risk patients. Additional analyses revealed that XPO1+Epithelial is influenced primarily by fibroblasts via ligand-receptor interactions, such as FN1-(ITGAV+ITGB1), and constitute a significant component of the TP-TME subtype. Moreover, XPO1+Epithelial interact with monocytes/macrophages, T/NK cells, and endothelial cells through ligand-receptor pairs, including MIF-(CD74+CXCR4), MIF-(CD74+CD44), and VEGFA-VEGFR1R2, respectively, thereby promoting the recruitment of immune-suppressive cells and angiogenesis. The ST-seq cohort treated with Tyrosine Kinase Inhibitors (TKIs) and Programmed Cell Death Protein 1 (PD-1) presented elevated levels of TP and TME risk subtype signature genes, as well as XPO1+Epithelial, T-cell, and endothelial cell infiltration in the treatment response group. Drug sensitivity analyses indicated that TP-TME high-risk subtypes, including sorafenib and pembrolizumab, were associated with sensitivity to multiple drugs. Further exploratory analyses revealed that CTLA4, PDCD1, and the cancer antigens MSLN, MUC1, EPCAM, and PROM1 presented significantly increase expression levels in the high-risk subtype group. This study constructed a robust prognostic model for HCC and identified novel subgroups at the single-cell level, potentially assisting in the assessment of prognostic risk for HCC patients and facilitating personalized drug therapy.

A hyperelastic constitutive model for soft elastomers considering the entanglement-dependent finite extensibility

In this paper, a novel hyperelastic constitutive model for soft elastomers is developed based on the concept of the tortuous tube. This model incorporates the finite extensibility of the polymer chain, the entanglement contribution to elasticity and the non-affine micro-to-macro scale transition in a unified way. To reflect the entanglement effect and its influence on the deformation of soft elastomers, the tortuous tube concept is introduced. The finite extensibility and conformational statistics of an entangled polymer chain in such a tortuous tube are clarified. By embedding the tortuous tube into the microsphere and employing the principle of minimum averaged free energy, a new non-affine scale transition rule is proposed to establish the relationship between the local deformation of a polymer chain at the microscopic scale and the overall deformation at the macroscopic scale. Based on the probability density function related to the conformational statistics, the Helmholtz free energy is established and further decoupled into a volumetric part and an isochoric part. The spatial Kirchhoff stress tensor and spatial elasticity tensor are derived from the newly established Helmholtz free energy. The proposed model is further implemented into the finite element program ABAQUS by writing a user-defined material subroutine. The prediction capability of the proposed model is verified by simulating the homogeneous and inhomogeneous deformations of soft elastomers under various loading modes, including uniaxial tension, uniaxial compression, pure shear, equi-biaxial tension, general biaxial tensile loadings, inflation and indentation. Moreover, the influence of entanglement concentration on the stretchability and stiffness of soft elastomers is predicted and discussed using the proposed model.

Metastable Oscillatory Modes as a Signature of Entropy Management in the Brain.

Entropy management, central to the Free Energy Principle, requires a process that temporarily shifts brain activity toward states of lower or higher entropy. Metastable synchronization is a process by which a system achieves entropy fluctuations by intermittently transitioning between states of collective order and disorder. Previous work has shown that collective oscillations, similar to those recorded from the brain, emerge spontaneously from weakly stable synchronization in critically coupled oscillator systems. However, direct evidence linking the formation of collective oscillations to entropy fluctuations is lacking. In this short communication, we demonstrate how the emergence of Metastable Oscillatory Modes (MOMs) is directly associated with a temporary reduction in entropy in the ongoing dynamics. We apply Shannon entropy to the distribution of eigenvalues of phase covariance over sliding time windows, capturing the temporal evolution of entropy at the level of the entire dynamical system. By demonstrating how the formation of MOMs impacts a system's entropy levels, we bridge theoretical works on the physics of coupled oscillators with the FEP framework, supporting the hypothesis that brain rhythms recorded experimentally are a signature of entropy management.

Towards Better Accuracy-Efficiency Trade-Offs: Dynamic Activity Inference via Collaborative Learning From Various Width-Resolution Configurations

Towards Better Accuracy-Efficiency Trade-Offs: Dynamic Activity Inference via Collaborative Learning From Various Width-Resolution Configurations

Affordances, mental health and psychiatry

This paper suggests that new directions in mental health and psychiatry are being driven by anomalies within epidemiology, mechanisms, and integration-based cognition. Three contrasting kinds of new work on mental disorder are critically discussed: enactive, ecological and active inference approaches, drawing from work by Varela, Gibson and Friston. These are detailed together with debates about their similarities and differences. All of these positions in fact make common use of ‘affordance’ as a core perception-action mechanism, suggesting promising options for new or enhanced psychiatric interventions. This leads to a critical review of the current state of knowledge on affordances, and their limitations, together with some suggestions as to how this approach might be applied in mental health and psychiatric practice.

PhosX: data-driven kinase activity inference from phosphoproteomics experiments.

The inference of kinase activity from phosphoproteomics data can point to causal mechanisms driving signalling processes and potential drug targets. Identifying the kinases whose change in activity explains the observed phosphorylation profiles, however, remains challenging, and constrained by the manually curated knowledge of kinase-substrate associations. Recently, experimentally determined substrate sequence specificities of human kinases have become available, but robust methods to exploit this new data for kinase activity inference are still missing. We present PhosX, a method to estimate differential kinase activity from phosphoproteomics data that combines state-of-the-art statistics in enrichment analysis with kinases' substrate sequence specificity information. Using a large phosphoproteomics dataset with known differentially regulated kinases we show that our method identifies upregulated and downregulated kinases by only relying on the input phosphopeptides' sequences and intensity changes. We find that PhosX outperforms the currently available approach for the same task, and performs better or similarly to state-of-the-art methods that rely on previously known kinase-substrate associations. We therefore recommend its use for data-driven kinase activity inference. PhosX is implemented in Python, open-source under the Apache-2.0 licence, and distributed on the Python Package Index. The code is available on GitHub (https://github.com/alussana/phosx).

Integration of Sense and Control for Uncertain Systems Based on Delayed Feedback Active Inference.

Asa result of the time lag in transmission, the data obtained by the sensor is delayed and does not reflect the state at the current moment. The effects of input delay are often overlooked in active inference (AIF), which may lead to significant deviations in state estimation and increased prediction errors, particularly when the system is subjected to a sudden external stimulus. In this paper, a theoretical framework of delayed feedback active inference (DAIF) is proposed to enhance the applicability of AIF to real systems. The probability model of DAIF is defined by incorporating a control distribution into that of AIF. The free energy of DAIF is defined as the sum of the quadratic state, sense, and control prediction error. A predicted state derived from previous states is defined and introduced as the expectation of the prior distribution of the real-time state. A proportional-integral (PI)-like control based on the predicted state is taken to be the expectation of DAIF preference control, whose gain coefficient is inversely proportional to the measurement accuracy variance. To adaptively compensate for external disturbances, a second-order inverse variance accuracy replaces the fixed sensory accuracy of preference control. The simulation results of the trajectory tracking control of a quadrotor unmanned aerial vehicle (UAV) show that DAIF performs better than AIF in state estimation and disturbance resistance.

Safeguarding User-Centric Privacy in Smart Homes

Internet of Things (IoT) devices have been increasingly deployed in smart homes to automatically monitor and control their environments. Unfortunately, extensive recent research has shown that on-path external adversaries can infer and further fingerprint people’s sensitive private information by analyzing IoT network traffic traces. In addition, most recent approaches that aim to defend against these malicious IoT traffic analytics cannot adequately protect user privacy with reasonable traffic overhead. In particular, these approaches often did not consider practical traffic reshaping limitations, user daily routine permitting, and user privacy protection preference in their design. To address these issues, we design a new low-cost, open source user-centric defense system—PrivacyGuard—that enables people to regain the privacy leakage control of their IoT devices while still permitting sophisticated IoT data analytics that is necessary for smart home automation. In essence, our approach employs intelligent deep convolutional generative adversarial network assisted IoT device traffic signature learning, long short-term memory based artificial traffic signature injection, and partial traffic reshaping to obfuscate private information that can be observed in IoT device traffic traces. We evaluate PrivacyGuard using IoT network traffic traces of 31 IoT devices from five smart homes and buildings. We find that PrivacyGuard can effectively prevent a wide range of state-of-the-art adversarial machine learning and deep learning based user in-home activity inference and fingerprinting attacks and help users achieve the balance between their IoT data utility and privacy preserving.

Bayesian Mechanics of Synaptic Learning Under the Free-Energy Principle.

The brain is a biological system comprising nerve cells and orchestrates its embodied agent's perception, behavior, and learning in dynamic environments. The free-energy principle (FEP) advocated by Karl Friston explicates the local, recurrent, and self-supervised cognitive dynamics of the brain's higher-order functions. In this study, we continue to refine the FEP through a physics-guided formulation; specifically, we apply our theory to synaptic learning by considering it an inference problem under the FEP and derive the governing equations, called Bayesian mechanics. Our study uncovers how the brain infers weight changes and postsynaptic activity, conditioned on the presynaptic input, by deploying generative models of the likelihood and prior belief. Consequently, we exemplify the synaptic efficacy in the brain with a simple model; in particular, we illustrate that the brain organizes an optimal trajectory in neural phase space during synaptic learning in continuous time, which variationally minimizes synaptic surprisal.

A hierarchical active inference model of spatial alternation tasks and the hippocampal-prefrontal circuit

Cognitive problem-solving benefits from cognitive maps aiding navigation and planning. Physical space navigation involves hippocampal (HC) allocentric codes, while abstract task space engages medial prefrontal cortex (mPFC) task-specific codes. Previous studies show that challenging tasks, like spatial alternation, require integrating these two types of maps. The disruption of the HC-mPFC circuit impairs performance. We propose a hierarchical active inference model clarifying how this circuit solves spatial interaction tasks by bridging physical and task-space maps. Simulations demonstrate that the model’s dual layers develop effective cognitive maps for physical and task space. The model solves spatial alternation tasks through reciprocal interactions between the two layers. Disrupting its communication impairs decision-making, which is consistent with empirical evidence. Additionally, the model adapts to switching between multiple alternation rules, providing a mechanistic explanation of how the HC-mPFC circuit supports spatial alternation tasks and the effects of disruption.

A sparse diverse-branch large kernel convolutional neural network for human activity recognition using wearables

During the past decade, large convolutional kernels have long been under the shadow of small convolutional kernels since the introduction of VGG backbone network. It always remains mysterious whether one can design pure convolutional neural network (CNN) while plugging larger kernels to model long-range dependency for human activity recognition (HAR), which has been rarely explored in previous literatures. In this paper, we revive the usage of larger kernels in the context of HAR and attempt to eliminate the performance gap between large kernels and small kernels by strategically applying a large receptive field, without incurring high memory and computational footprints. Built on two recipes, i.e., Diverse-Branch and Dynamic Sparsity, we design a pure CNN architecture named SLK-Net for activity recognition, which is equipped with sparse diverse-branch larger kernels. To validate the effectiveness of our approach, we perform a series of extensive experiments on four public benchmarks including UCI-HAR, WISDM, UniMiB-SHAR and USC-HAD, which show that large kernels can benefit its ability to capture long-range dependency and consistently beat state-of-the-art small-kernel counterparts across a wide range of activity classification tasks. Real activity inference latency is measured on a mobile device, which reveals that such sparse diverse-branch kernels can lead to inference speedup than vanilla large kernels. We hope this work may further inspire relevant CNN-based studies in the HAR community.

Active Inference Models of AV Takeovers: Relating Model Parameters to Trust, Situation Awareness, and Fatigue.

Our objectives were to assess the efficacy of active inference models for capturing driver takeovers from automated vehicles and to evaluate the links between model parameters and self-reported cognitive fatigue, trust, and situation awareness. Control transitions between human drivers and automation pose a substantial safety and performance risk. Models of driver behavior that predict these transitions from data are a critical tool for designing safer, human-centered, systems but current models do not sufficiently account for human factors. Active inference theory is a promising approach to integrate human factors because of its grounding in cognition and translation to a quantitative modeling framework. We used data from a driving simulation to develop an active inference model of takeover performance. After validating the model's predictions, we used Bayesian regression with a spike and slab prior to assess substantial correlations between model parameters and self-reported trust, situation awareness, fatigue, and demographic factors. The model accurately captured driving takeover times. The regression results showed that increases in cognitive fatigue were associated with increased uncertainty about the need to takeover, attributable to mapping observations to environmental states. Higher situation awareness was correlated with a more precise understanding of the environment and state transitions. Higher trust was associated with increased variance in environmental conditions associated with environmental states. The results align with prior theory on trust and active inference and provide a critical connection between complex driver states and interpretable model parameters. The active inference framework can be used in the testing and validation of automated vehicle technology to calibrate design parameters to ensure safety.

Active inference and deep generative modeling for cognitive ultrasound.

Ultrasound has the unique potential to offer access to medical imaging to anyone, everywhere. Devices have become ultra-portable and cost-effective, akin to the stethoscope. Nevertheless, and despite many advances, ultrasound image quality and diagnostic efficacy are still highly operator- and patient-dependent. In difficult-to-image patients, image quality is often insufficient for reliable diagnosis. In this paper, we put forth the idea that ultrasound imaging systems can be recast as information-seeking agents that engage in reciprocal interactions with their anatomical environment. Such agents autonomously adapt their transmit-receive sequences to fully personalize imaging and actively maximize information gain in-situ. To that end, we will show that the sequence of pulse-echo experiments that an ultrasound system performs can be interpreted as a perception-action loop: the action is the data acquisition, probing tissue with acoustic waves and recording reflections at the detection array, and perception is the inference of the anatomical and or functional state, potentially including associated diagnostic quantities. We then equip systems with a mechanism to actively reduce uncertainty and maximize diagnostic value across a sequence of experiments, treating action and perception jointly using Bayesian inference given generative models of the environment and action-conditional pulse-echo observations. Since the representation capacity of the generative models dictates both the quality of inferred anatomical states and the effectiveness of inferred sequences of future imaging actions, we will be greatly leveraging the enormous advances in deep generative modelling (generative AI), that are currently disrupting many fields and society at large. Finally, we show some examples of cognitive, closed-loop, ultrasound systems that perform active beamsteering and adaptive scanline selection, based on deep generative models that track anatomical belief states.